NFS server not responding

[ricky_8]

Hi ALL,

This is the message i get on my mentorgraphics client application .We have mentor server and clients for users to access, this is a new setup and in testing phase.
The NFS connections are right we can access all the data exported from servers on this clients but sometimes i get the NFS server not responding messages and application does not start . If i check access to the nfs filesystems they show no problem of access.
What might be the problem isit related to NFS patching or mentor graphics application?.
In syslogs i see "NFS server not responding "and on next line "NFS server is ok"

Has anyone encountered such a problem?Any help?

[ricky_8]

Would like to add some more information. All along ping works from server to client vice versa.nslookup is able to resolve the hostname for client and server both on both ends.though i am using IP address for mounting the Filesystem so no issue of name resolution.

Other thing is suddenly it starts working fine with no clue what made it work.

[Michael Tully]

You could be encountering some form of lan traffic problem with NFS. You may wish to increase the number of background 'nfsd' daemons.
This is done in the /etc/rc.config.d/nfsconf file.
These are the entries that may need to be increased, e.g.
NUM_NFSD=16
NUM_NFSIOD=4

BIOD helps the reads, while NFSD helps the writes. Adding more BIOD does
nothing for you if you are writing your data to the NFS server. BIODs
help quite a bit on reads, depending on the number of processes and
users that do concurrent reads. 4 to 16 are typical numbers. You could
use 16 for NFSD, 4 for BIOD.

[S.K. Chan]

I have a previous reply that has some troubleshooting tips for NFS (10.x) which you may find useful. See..
http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0xae40ec08252fd611abd50090277a778c,00.html
It sounds like your problem is intermittent .. I do occasionally see in my /etc/dmesg output (of a couple of my servers) having the same error message but have not gotten around to troubleshoot it because it did not last very long. You would definately want to make sure you are uptodate on NFS patches.

[Michael Tully]

You could be encountering some form of lan traffic problem with NFS. You may wish to increase the number of background 'nfsd' daemons.
This is done in the /etc/rc.config.d/nfsconf file.
These are the entries that may need to be increased, e.g.
NUM_NFSD=16
NUM_NFSIOD=4

BIOD helps the reads, while NFSD helps the writes. Adding more BIOD does
nothing for you if you are writing your data to the NFS server. BIODs
help quite a bit on reads, depending on the number of processes and
users that do concurrent reads. 4 to 16 are typical numbers. You could
use 16 for NFSD, 4 for BIOD.

[Michael Tully]

You could be encountering some form of lan traffic problem with NFS. You may wish to increase the number of background 'nfsd' daemons.
This is done in the /etc/rc.config.d/nfsconf file.
These are the entries that may need to be increased, e.g.
NUM_NFSD=16
NUM_NFSIOD=4

BIOD helps the reads, while NFSD helps the writes. Adding more BIOD does
nothing for you if you are writing your data to the NFS server. BIODs
help quite a bit on reads, depending on the number of processes and
users that do concurrent reads. 4 to 16 are typical numbers. You could
use 16 for NFSD, 4 for BIOD.

[Michael Tully]

You could be encountering some form of lan traffic problem with NFS. You may wish to increase the number of background 'nfsd' daemons.
This is done in the /etc/rc.config.d/nfsconf file.
These are the entries that may need to be increased, e.g.
NUM_NFSD=16
NUM_NFSIOD=4

BIOD helps the reads, while NFSD helps the writes. Adding more BIOD does
nothing for you if you are writing your data to the NFS server. BIODs
help quite a bit on reads, depending on the number of processes and
users that do concurrent reads. 4 to 16 are typical numbers. You could
use 16 for NFSD, 4 for BIOD.

[Michael Tully]

You could be encountering some form of lan traffic problem with NFS. You may wish to increase the number of background 'nfsd' daemons.
This is done in the /etc/rc.config.d/nfsconf file.
These are the entries that may need to be increased, e.g.
NUM_NFSD=16
NUM_NFSIOD=4

BIOD helps the reads, while NFSD helps the writes. Adding more BIOD does
nothing for you if you are writing your data to the NFS server. BIODs
help quite a bit on reads, depending on the number of processes and
users that do concurrent reads. 4 to 16 are typical numbers. You could
use 16 for NFSD, 4 for BIOD.

[ricky_8]

In NFS diagnostic guide it says following
nfsstat -rc
Client rpc:
calls badcalls retrans badxid timeout wait newcred
3768121 0 0 9 122 0 0

=> If timeout and retrans are high but badxid is close to zero, packets are
being dropped before they get to the NFS server.
Action
------
o Reduce values of wsize and rsize mount options to 4096 or 2048 on clients.


But how to change the rsize and wisze values and how to check their existing sizes?

[john korterman]

Hi Ricky,
A timeout value of 122 and a retrans value of 0 can hardly be called high ??? at least not both of them; probably no background for changing wsize and rsize.
Are you familiar with the attached trouble-shooting document?
http://www4.itrc.hp.com/service/cki/docDisplay.do?docLocale=en_US&docId=200000066581018


regards,
John K.

[ricky_8]

Well those values were from the document.

Here are my client values :

Client rpc:
Connection oriented:
calls badcalls badxids timeouts newcreds badverfs timers
0 0 0 0 0 0 0
cantconn nomem interrupts
0 0 0
Connectionless oriented:
calls badcalls retrans badxids timeouts waits newcreds
36401 100 367 0 444 0 0
badverfs timers toobig nomem cantsend bufulocks
0 60 0 0 0 0

I would like to know how to change wsize and rsize.

[john korterman · Accepted solution]

Hi again,
then try this document:

http://www4.itrc.hp.com/service/cki/docDisplay.do?docLocale=en_US&docId=200000062937497

regards,
John K.

[ricky_8]

John,

I was not able to open this doc, can you help me to attach it .

Thanxs a lot

[john korterman]

Hi,
hope it is readable..

regards,
John K.

[ricky_8]

Thanxs a lot john,

I was able to see it and seems to be working ... thats very use doc .

Thanxs a gain.

[ricky_8]

Thanxs a lot john,

I was able to see it and seems to be working ... thats very useful doc .

Thanxs a gain.

[ricky_8]

Thanxs a lot john,

I was able to see it and seems to be working ... thats very useful doc .

Thanxs a gain